@dgelessus

35 Followers
118 Following
601 Posts

person who computers. Casual disassembler of data formats. Standards and specifications enjoyer. Using Python for everything, even if it's a bad idea.

(really though, I'm just here to post silly puns)

^_^

GitHubhttps://github.com/dgelessus
Bluesky bridgehttps://bsky.app/profile/dgelessus.mastodon.social.ap.brid.gy

My current solution: Have a global "loose threads" object that gets to own all the threads that have no well-defined scope. Before the application exits, it waits for all of those loose threads to finish.

So basically I reconstructed Java's thread behavior by hand. Not satisfying, but it works 🤷‍♀️

Web developers who break Ctrl/Cmd+click for "open in new tab" should be forced to work with a one-button mouse until they atone for their sins
Haven't decided yet if I like this restriction or not. It *is* causing me a headache, but it might be the good kind of headache, because it's forcing me to decide which part of my code owns those would-be loose threads (and is thus responsible for waiting for them to finish).
This restriction is especially odd because Qt supports cross-thread communication (via signals/slots) in a way that either side can safely "disappear" at any time without risking dangling pointers or anything. Would be perfect for detached threads that aren't coordinated with their parent thread!
Qt's QThread insists that if you use a QThread object to start a thread, then you cannot delete the QThread object as long as that thread is still running. This is causing me quite the headache, because I'd really like to just let a thread run independently without caring about when it finishes...
python3 -m pip install --edible .

A few weeks ago, someone on here posted a link to a very long, detailed, interesting article on modern CPU design. I still haven't finished reading it, and meanwhile my reading list keeps getting longer.

That is to say, I'm encountering a pipeline stall caused by modern CPU design and slow reads

Highly disappointed to find out that Oracle's Greek offices are in Athens and not in Delphi

https://www.oracle.com/europe/corporate/contact/field-offices/#greece

idea: "booter service", except it's not DDoS-for-hire - it's just a guy who will make you any kind of boot media you need, and can help you flash working firmware onto bricked mobile devices